Michael and Yuhang

Following the method provided in Marco Thesis, we made a phase noise measurement for AFG3251. We take signal from DDS as a reference signal, since it was reported from Marco thesis that this is a stable signal source. We will check the phase noise of DDS later.

The measurement was to use DDS and AFG3251 to generate 20MHz signal and acquire their beatnote after low pass. We tried to synchronize them with another DDS channel to provide a 10MHz signal to the 'external reference input' channel of AFG3251 (Fig. 1 and 2). However, the 20MHz between them cannot be synchronized. In the end, we tune DDS to be 20.00073506 MHz to have beatnote signal as flat as possible. In the end, there was still some very low frequency component was not removed (Fig. 4). Then we took measurement of beatnote with different frequency band using network analyzer.

To calibrate the measured beatnote signal, we increased DDS signal frequency by 1kHz. Note that we used DDS channel with amplitude reduced by a factor of 2 to avoid saturation. And we found pk-pk beatnote is 0.316V. Therefore, the measured beatnote spectrum need to be divided by (0.316/2).

The phase noise is shown in Fig. 3. The integrated phase noise from high frequency to 10Hz is 399urad, which is slightly higher than the measured DDS phase noise (reported to be 117urad from Marco thesis).